(269g) Determination of H2O-Solvated Cationic Fe(III) Coordination Geometry in Fe-SSZ-13 Using Wavefunction Coupled-Cluster Parameterized Hybrid Density Functional Theory

Zeolites with Fe dispersed as cations have been widely reported as effective industrial and environmental catalysts for a wide range of reactions. Identifying the state of cationic Fe centers is important for rationalization of the corresponding catalytic functions that allows strategy development for further material optimization. Among various geometrical or electronic descriptors, Fe coordination geometry is a geometrical fingerprint that allows identification of ligands and adsorbates attached to Fe under a specific condition, which can be inferred from spectroscopic spectra (e.g. UV-Vis, Extended X-Ray Absorption Fine Structure, etc.), but generally require peak assignment or model fitting, so that conclusions greatly depend on the type of spectroscopy used and individual interpretation of the spectra. An alternative is to explicitly simulate Fe coordination from first-principles using the major workhorse for zeolite modeling, density functional theory (DFT). One of the technical challenges, however, is the accuracy of DFT in probing transition metal coordination chemistry. Recent literature suggests hybrid DFT with parameterized Hartree-Fock exact exchange (HF) incorporation is required to qualitatively reproduce transition metal (e.g. Cu(II)) coordination geometry preference predicted by the benchmarking wavefunction method.

In this work, we focus on H2O-solvated monomeric Fe(III) complexes, which have been shown to be the dominant species in zeolites under ambient conditions. We first benchmarked plane-wave supercell GGA and hybrid HSE functionals against local basis set wavefunction coupled-cluster method to determine optimal HF fraction for modeling coordination of H2O-solvated Fe(III) clusters. We observed that both GGA and HSE with insufficient HF lead to qualitatively erroneous conclusions, highlighting the importance of theory benchmarking. Further, we applied the optimized HSE functionals for zeolite systems. We used SSZ-13 as a model zeolite framework as it has only one symmetry-distinct tetrahedral-site (T-site), which avoids the complexity of decoupling T-site preferences for cationic Fe siting. We examined H2O-solvated Fe(III) charged-compensated by either single or paired Al sites in SSZ-13 and observed the trigonal bipyramidal coordination preference for zeolitic Fe(III) clusters, which is different from its octahedral preference in the aqueous phase. We further performed ab initio molecular dynamics on the preferentially-coordinated complexes for simulations of the radial distribution function (RDF) between Fe and the surrounding heavy atoms (O, Si, and Al). The simulated RDFs are qualitatively distinguishable between single and paired Al sites, potentially allow comparison with EXAFS for Fe center identification at the molecular level. Our findings also highlight a caveat in general of DFT accuracy during theoretical interrogation of transition metal cation coordination relevant problems, e.g. predicting preferred coordination geometry, computing energetics with a change of coordination geometry and/or number between intermediates, etc.